John Hudson responded:
> At 17:21 5/7/2002, Kenneth Whistler wrote:
>
> > > The missing character is called "ga" (attached). It looks like Kaf
> > (06A9) with a dot on top.
> >
> >The intended letter for Jawi ga is:
> >
> >U+06AC ARABIC LETTER KAF WITH DOT ABOVE
>
> The problem with this is that the Jawi ga, as far as I know, does not use
> the isolated and final form based on lam with miniature swash kaf (first
> row of attached GIF), but uses the 'zigzag' form for all four basic
> presentation glyphs (second row).
I concur with the glyph analysis, which also matches Alan Kaye's
discussion of Jawi in Daniels and Bright.
However, I also think it is quite clear that U+06AC ARABIC LETTER KAF
WITH DOT ABOVE is the one that was intentionally encoded for the Jawi
ga character. This has been marked as an "old Malay" letter ever since
Unicode 1.0.
The error would seem, rather, to be in choosing an inappropriate
representative glyph for the charts, and in the shaping class
assigned to the character in ArabicShaping.txt:
06A9; OPEN KAF; D; GAF
06AB; KAF WITH RING; D; GAF
06AC; KAF WITH DOT ABOVE; D; KAF
06AD; KAF WITH 3 DOTS ABOVE; D; KAF
06AE; KAF WITH 3 DOTS BELOW; D; KAF
06AF; GAF; D; GAF
I'd say, as for the Pashto Kaf with ring, that the appropriate shaping
class should be "GAF", not "KAF", in this file.
Let's pass this by other Arabist specialists in the UTC before
taking further action.
--Ken
> Linotype always treated the Jawi ga as a
> variant of the gaf, with the dot replacing the stroke, and I've been
> presuming Jawi encoding based on this. It doesn't really matter, since the
> Unicode fonts we've built contain U+06AC as well as variant forms for
> U+06AF, so either encoding is supported, but perhaps Bustamam can confirm
> which isolated and final forms are preferred for writing Malay?
>
> John Hudson
This archive was generated by hypermail 2.1.2 : Tue May 07 2002 - 22:31:23 EDT